Vous êtes sur la page 1sur 20


discussions, stats, and author profiles for this publication at:


Interval training for performance: A

scientific and empirical practice:
Special recommendations for middle-
and long-distance running. Part II:
Anaerobic interval training

Article in Sports Medicine March 2001

DOI: 10.2165/00007256-200131020-00001 Source: PubMed


94 1,500

1 author:

Veronique Louise Billat

Universit d'vry-Val-d'Essonne


Some of the authors of this publication are also working on these related projects:

a minimalist training View project

Available from: Veronique Louise Billat

Retrieved on: 19 October 2016
LEADING ARTICLE Sports Med 2001; 31 (1): 13-31

Adis International Limited. All rights reserved.

Interval Training for Performance:

A Scientific and Empirical Practice
Special Recommendations for Middle- and Long-Distance
Running. Part I: Aerobic Interval Training
L. Vronique Billat
Faculty of Sport Science, University Lille, Lille, France

Abstract This article traces the history of scientific and empirical interval training.
Scientific research has shed some light on the choice of intensity, work duration
and rest periods in so-called interval training. Interval training involves repeated
short to long bouts of rather high intensity exercise (equal or superior to maximal
lactate steady-state velocity) interspersed with recovery periods (light exercise
or rest). Interval training was first described by Reindell and Roskamm and was
popularised in the 1950s by the Olympic champion, Emil Zatopek.
Since then middle- and long- distance runners have used this technique to train
at velocities close to their own specific competition velocity. In fact, trainers have
used specific velocities from 800 to 5000m to calibrate interval training without
taking into account physiological markers. However, outside of the competition
season it seems better to refer to the velocities associated with particular physi-
ological responses in the range from maximal lactate steady state to the absolute
maximal velocity. The range of velocities used in a race must be taken into con-
sideration, since even world records are not run at a constant pace.

1. Definition and Characteristics of plaining why the accepted training practices result
Interval Training in an enhanced performance. Far less frequently have
scientific breakthroughs, arising from the labora-
Training can be defined as the systematic and reg- tory, precipitated major changes in accepted train-
ular participation in exercise to enhance sports per- ing practices.[1] However, scientific research has
formance. Performance, especially for sports based shed some light on the choice of intensity, work du-
on locomotion (action to move from one point to ration and rest periods in so-called interval train-
another), can be a time to cover a distance (14 min- ing. Interval training involves repeated short to
utes over 5000m is a performance) or a distance long bouts of rather high intensity exercise (equal
covered in a time (21km in an hour of running). If or superior to maximal lactate steady-state veloc-
one considers the velocity-time relationship, the pur- ity) interspersed with recovery periods (light exer-
pose of training is to shift the curve to the right; to cise or rest). Interval training was first described,
be able to run faster over the same distance or to in a scientific journal, by Reindell and Roskamm,[2]
run a longer time at a given velocity. To date, sports and Reindell et al.,[3] and was popularised in the
scientists have focused much of their effort on ex- 1950s by the Olympic champion, Emil Zatopek.
14 Billat

Since then middle- and long-distance runners have velocity is 75% vVO2max, the amplitude is: 100
used this technique to train at velocities close to 75/75 = 33%;
their own specific competition velocity. (iv) the duration and the distances run at high
Indeed, interval training has a very long tradi- and low velocities.
tion. Often cross-country running (and skiing) was To appreciate both the immediate and the long
included in the training with running (skiing) on term effects of interval training programmes, mod-
the flat, uphills, and downhills, and it was termed ern, technological progress has provided devices
natural interval training. In this kind of interval allowing field measurements of the physiological
training the athletes often guided their speed with responses of athletes during interval running. The
a stop-watch (e.g. the Swedish runner, Gunder Hgg different types of interval training which have been
who in early 1940 achieved 15 individual world used to improve aerobic and anaerobic capacity are
presented in table I.
records in middle- and long- distance running). The
problem was that these training programmes were
not published in scientific journals. Interval train- 2. The Pioneers of Interval Training:
ing, including measurements of heart rate, oxygen Physiologists, Trainers and Runners
. .
uptake (VO2) and blood lactate levels, was applied In 1910, it was possible to measure VO2 during
to the successful Swedish cross-country skiers in exercise but no athletes were tested for training im-
the 1950s. At that time, it was already known that provement. However, in 1912, the 10 000m Olym-
for a given distance, which at maximal speed took pic championship runner, Hannes Kolehmainen
4.0 minutes to cover, could be prolonged to 4 min- (Finland), had already used interval training at the
utes 30 seconds and still load the oxygen transport specific 10km pace. He had trained using 5 to 10
system to maximum without much less accumula- repetitions of 3 minutes 5 seconds every 1000m (19
tion of lactate. Those intervals could be repeated km/h). 80 years later the 10km specific interval train-
many times before fatigue. Scientific publications ing is run at 22.7 km/h.
of physiological data were published 2 decades later. During the 1920s and 1930s, at a time when Hill[5]
If we take, for instance, an interval training of 3 had invented the concept of VO2max and oxygen
minutes at 100% of the minimal velocity associated deficit to explain the shape of the velocity-time re-
. lationship, the great Finnish runner, Pavoo Nurmi
with the maximal oxygen consumption (vVO2max)
determined in an incremental test interspersed with (who ran the 5000m in 14 minutes 36 seconds at
. 20.6 km/h), introduced short interval training at an
3 minutes at 50% vVO2max, interval training char-
intensity superior to a specific velocity such as 6
acteristics as defined by Saltin et al.[4] will have the
400m in 60 seconds at 24 km/h inside a slow run
following characteristics:
of 10 to 20km in the woods.
(i) the intensity is defined as the average power
After the second world war, interval training be-
output; for the interval training described above,
came a widespread training method used by Eu-
the average intensity is equal to (100 + 50)/2 = 75%
. ropean runners. Emil Zatopek (Czechoslovakia, tri-
vVO2max [about 75% of maximal oxygen uptake
. ple gold medallist in 1952 in 5000, 10 000m and
(VO2max)]; Marathon events), Gordon Pirie (UK, 3000m in
(ii) the time-ratio for the high and low exercise 7 minutes 57 seconds in 1960), Sigfried Hermann
duration; for the interval training described above, (Germany, 800m in 1 minute 48 seconds, and 1500m
the time ratio 3/3 = 1; in 3 minutes 40.9 seconds) trained by Toni Nett,
(iii) the amplitude is the ratio of the difference Roger Moens (Belgium), and Vladimir Kutz (USSR,
between the intensity of the different periods (heavy 5000m in 13 minutes 35.0 seconds) all used inter-
or recovery run) with the average velocity; for the val training. The most famous athlete to use inter-
interval training described above, since the average val training was Emil Zatopek who initiated short

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
Aerobic Interval Training and Performance 15

Table I. Classification of the different types of interval training according to the specific velocities of a race, the time limit at these velocities
and physiological velocities: the velocity at maximal oxygen uptake (VO2max), the critical velocity (i.e. the asymptote of the velocity-time limit
relationship), and the velocity at maximal lactate steady state
Intensity Physiological Time limit at Time spent Maximal blood Aerobic Anaerobic interval Aerobic interval
. .
(% vVO2max) and competition this velocity at VO2max lactate level metabolism training training
velocity (min) (min) (mmol/L) to energy (%)
115-130 v1000m; 3-2 2-1 15-18 75-65 6 30 sec; R = 30 20 10 sec;
v800m sec (rest); 60 sec, R = 10 sec (rest)
45 sec, 30 sec, 45
sec, 60 sec; R = 5
min (rest)
105-115 vmiles; 6-4 4-2 13-15 85-80 6 1 min; R = 3 min 15 15 sec;
v1500m (rest); 3 500m at R = 15 sec at 50%
v1500m; R = 3 min vVO2max
100-105 vVO2max; 8-6 5-4 11-13 90-85 3 1000m at 20 15 sec;
v3000m v3000m; R = 3 min R = 15 sec at 50%
(rest) vVO2max
95-100 v5000m 15-8 10-5 9-11 95-90 5 1000m at 25 15 sec;
v5000m; R = 3 min R = 15 sec at 50%
(rest) vVO2max; 6 3 min;
R = 3 min 50% vVO2max
90-95 v10 000m and 30-15 1-10 7-9 97.0 3 3000m at
critical velocity v10 000m; R = 3 min
85-90 Velocity for 60-30 0 5-7 98.0 2 20 min; R = 3 min
record of the at 70% vVO2max
80-85 Maximal lactate 80-60 0 3-5 99.0 -2 30 min; R = 3 min
steady state at 70% vVO2max
75-80 Marathon 150-80 0 3-3.5 99.9 2 15km; R = 1km at
velocity 70% vVO2max
R = recovery between series (i.e. set of several repetitions); vVO2max = velocity at maximal oxygen uptake; vxm = average velocity over x

interval training at low amplitudes and running at and the last series faster [28 seconds, i.e. 105% of
the critical velocity. His critical velocity, calculated the average velocity over 1500m (v1500m) with
from his personal best in 3 to 10km events accord- the last run in 25 seconds, i.e. 118% of v1500m].
ing to Ettema,[6] was about equal to 85% vVO2max, If one looks at the variation in pace per 300m,
that is, 20 km/h, or 1 minute 12 seconds in 400m even for the worlds records, to choose the inten-
or lower at (probably) his maximal blood lactate sity and duration of interval training, it seems ap-
steady state. Indeed, he repeated up to 100 400m propriate to train within the range of the velocity
repetitions per day, interspersed by 200m of recov- of the race. Therefore, trainers have always started
ery run at a pace close to that of hard work. from the requirements of the race, and chosen val-
Toni Nett (in Reindell et al.[3]) reported the spe- ues appropriate to the best performance of the run-
cific interval training programme of Sigfried Her- ner. However, scientists can provide information
mann for the 1500m (best performance 3 minutes regarding the physiological responses of the ath-
40 seconds): 4 (6 200m) with a rest of 50 to 60 lete during such interval training and it could be
seconds between runs and of 8 minutes between the possible to determine the effect of active or passive
series. The first series was run slower (30 seconds, i.e. pause. Unfortunately, interval training studied in
24 km/h at 98% of the average velocity over 1500m) the laboratory, even on the treadmill, has been cal-

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
16 Billat

ibrated with reference to vVO2max and not to the Table II. Summary of a typical day of interval training for Vladimir
Kutz (5000m in 13min 35.0sec, v5000m = 22 km/h)
best performance of the runners. The use of the criti-
1. 30min of jogging
cal velocity calculated from the slope of the distance-
2. 8 100m of acceleration (14sec)
time limit from the best performance over 3000 to
3. Stretching
10 000m could be a way to calibrate interval train- 4. 20 200m in 28-29sec (118% of v5000m) with recovery trot
ing taking into account the performance instead of
. 5. 4 400m (68sec, 96% of v5000m) with recovery trot
a physiological velocity, such as vVO 2max.[7] 6. 15min of light running where the runner has to focus on
By examining the interval training performed flexible style
Recovery times were not specified
during a week by Vladimir Kutz [the average ve-
On Friday the interval training programme was much more
locity over 5000m (v5000m) = 22 km/h], we can consecutive but it was added just before the stretching, 5
conclude that he practised many series separated 120-150m at maximal velocity and 20 400m (60-68sec)
by gymnastics. Nowadays, runners train twice a It provided:
day rather than performing such a long programme. 1. 30min of jogging
A typical day of training for Vladimir Kutz is pre- 2. 5 120-150m at maximal velocity
3. Stretching
sented in table II. Kutz performed interval training
4. 5 200m in 28-29sec (118% of v5000m) with recovery trot
in parks or woods. It should be pointed out that in 5. 20 400m (60-80sec in 96-109% of v5000m) with recovery
table II the interval training used by Vladimir Kutz trot
was composed of very different distances and ve- 6. 5 200m in 28-29sec (118% of v5000m) with recovery trot
locities, even within the same training session. 7. 15min of light running where the runner has to focus on
flexible style
In the fifties, Franz Stamfl who trained Roger
8. 15min of stretching
Bannister, the first sub-4-minute miler, employ- v5000m = average velocity over 5000m.
ed interval training in varying forms (aerobic, an-
aerobic; table I), 5 days per week almost all year
round. He preferred his athletes to run fewer miles .
100% of vVO2max: 10-second runs interspersed with
but he insisted on the quality of work rather than .
10 seconds of complete rest, since VO2 reached
the quantity although, once per week, they did per- .
VO2max with a low blood lactate accumulation. In
form Fartlek runs of 60 to 90 minutes duration.[1] 1960 the first study[9] describing the metabolic re-
However, Fartlek runs were actually previously in- sponse during interval training with particularly short
vented by the Swedish coach Gsta Holmr in the periods from 5 to 30 seconds, was published. This
1930s. was remarkable considering the absence of auto-
matic methods for measuring VO2. They reported
2.1 The 1960s .
that a runner (the individual BS, with a VO2max =
The sixties were the years of the first scientific 5.6 L/min, 67 ml/min/kg), performing very short
intermittent runs (15 seconds) of alternating brief
studies on interval training. In 1960, the pioneer .
Swedish physiologist Per Olf Astrand developed heavy intensity repetitions at 100% of vVO2max with
long interval training at a velocity between the crit- complete rest (15 seconds), sustained this exercise
. . for 30 minutes with a low level of blood lactate (2.3
ical velocity and vVO2max (90 to 95% vVO2max) .
[table I]. These 3 minutes run at about 90 to 92% mmol/L). Moreover, this runner reached VO2max at
. . the end of the exercise. Using shorter pauses (10
of vVO2max elicited VO2max in the last repetitions, .
despite the complete rest in between. Astrand et seconds) BS reached 95% of VO2max by the 18th
al.[8] considered that this was one of the best forms minute and the end blood lactate level stayed rather
of interval training to improve VO2max since all car- low, that is, 5.6 mmol/L. However, because of pas-
diorespiratory parameters were at their maximum. sive recovery this value fluctuated between 89 and
From the same group of researchers, Christensen 95% of VO2max. With shorter work and pauses (5
et al.[9] proposed very short interval training run at seconds, 5 seconds), this individual reached only

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
Aerobic Interval Training and Performance 17

81% of VO2max but this value was more constant work output and the fitness of the individual and
because of the very brief pause. The end blood lac- was exceptionally resistant to changes in the men-
tate level was just above 2 mmol/L (2.5 mmol/L), tal state of the trained individual. They pointed out
similar to that seen with very short intermittent runs the fact that in 4 minutes (for this individual and at
and pauses (15 seconds, 15 seconds). this work rate) the athlete reached VO2max. This
In the 1960s, the first studies by the Astrand and fact is of importance when the purpose is to choose
Christensen group examining the immediate and the duration of interval training. Indeed, about 50%
long term effects on metabolism were published. of the time at this work rate is required to reach
. .
In the first study,[10] they compared the same work VO2max. If the goal is to elicit VO2max at the first
performed at the same power output (360W and repetition, the duration must be equal to at least
almost 98% of power output, i.e. pVO2max) but with 50% of the time limit.[13] Astrand and Saltin, in
different work durations (30 seconds, 1, 2 and 3 1961,[14] demonstrated that because of the acceler-
minutes). The continuous time limit of this heavy ation in oxygen kinetics with high work rate, VO2max
exercise was equal to 9 minutes. It was found that was reached and maintained (about half of the time
when the cycling exercise was split into short pe- limit) for exercises lasting between 2 and 8 min-
riods of work and rest it was transformed into a utes. Actually, oxygen peak could, after warming
submaximal load on both circulation and respira- up, be attained within 1 minute. More than 10 years
tion (63% VO2max, blood lactate level 2 mmol/L) later, the same team,[15] did biopsies after each of the
and, hence, was well tolerated during 1 hour. With 5 repetitions of 1 minute of exercise at about 120%
longer periods (2 or 3 minutes duration) the work of pVO2max followed by 5 minutes of rest. Creatine
output became close to the upper limit of perfor- phosphate (CP) was progressively depleted after
mance and could be fulfilled only with the utmost each repetition and muscle lactate level reached its
strain (blood lactate of 16.6 mmol/L, and VO2max highest value (23 mmol/L/kg wet muscle), at the
at 100%). From a practical point of view the au- first work period. Blood lactate levels took more
thors stressed that by choosing longer periods, for time to increase, but reached 20 mmol/L by the
example 2 or 3 minutes, one could obtain a higher third repetition. Hermansen[16] had previously re-
training effect on cardiorespiratory function. More- ported very high levels of blood lactate (30 mmol/L)
over, to explain the low lactic acid values during using the same kind of interval training.
the short periods of work and rest it was proposed At the end of the sixties, the American group of
that the myoglobin functions as an oxygen store Fox et al.[17] focused on interval training in a mil-
during short spells of heavy muscular work.[8] itary context.[17,18] They compared metabolic en-
In another paper published in the same vol- ergy sources during continuous and interval run-
ume,[11] the same group hypothesised that myoglo- ning at the same rate. Moreover, they compared the
bin would represent an oxygen store which is used physiological response for a recovery run (60% of
during the initial phase of work before respiration vVO2max) or passive complete rest. They emphasised
and circulation are able to reach the values which that coaches had succeeded in improving the per-
correspond to the actual oxygen demand. This ox- formance of highly trained athletes using the inter-
ygen store was calculated to be equal to 0.43L, val training method. These investigators explained
which represents about 10% of the maximal accu- this by the fact that there is a slower accumulation
mulated oxygen deficit obtained in an all-out exer- of lactic acid, and therefore a delay in the onset of
cise of 2 minutes.[12] fatigue. This results from the replenishment and sub-
These investigators also described the oxygen sequent reutilisation of part of the phosphagen re-
kinetics in the all-out exercise at pVO2max (time serves which enable the athlete to accomplish large
limit: 9 minutes). They emphasised that the time quantities of work (distance) at very high intensi-
course of oxygen increase was dependent on the ties. However, these investigators advised coaches

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
18 Billat

to alternate the work intervals with rest, rather than papers, no VO2max values were routinely measured
running, to be able to restore phosphocreatine re- for training advice purposes.
serves. They did not control the amount of time
spent at VO2max, but just the work performed at 2.2 The 1970s and 1980s
high intensity, which supports the training concept .
During the seventies VO2max began to be sys-
of Tim Noakes,[19] but is in contrast to that of Jack
tematically measured in athletes, and in the eighties
the lactate threshold was measured. East German
During the same period, the New Zealand trainer
physiologists, such as Alois Mader, determined a
Arthur Lydiard (who trained Peter Snell, double
blood lactate threshold at 4 mmol/L using stages of
gold medallist in 1960) also developed a very short
. constant velocity, 5 minutes long (for review see
interval training method run at 100% of vVO2max. Billat[25]).
The duration of this short interval training, run at
. The eighties were years of exceptional runners
100% of vVO2max, was 10- to 15-second runs with such as Sebastian Coe (800 to 1500m). Coe was
pauses of the same duration run at 30 to 40% of
. trained by his father Peter who was very inspired
VO2max. We shall see in section 3 that this proce-
. by scientific methods. Sebastian Coe performed aer-
dure of short interval training at 100% vVO2max obic and anaerobic interval training as well as cir-
with active pauses allows the runner to stay a very cuit training for strength and power improvement.
long time at VO2max. Indeed, this has recently been North African runners, such as Said Aouita the great
demonstrated on the track using a portable breath middle-distance runner (who held the world records
by breath gas exchange analyser (K4b2, Cosmed, for the 1500 to 5000m) used interval training ses-
Italy). In addition to this short interval training, sions with different velocities. In the same interval
Lydiards athletes (even the middle-distance run- training session, he ran at velocities from the max-
ners) regularly performed training runs of 2 hours imal lactate steady-state velocity to v5000 (94%
duration (100 miles per week) as the Kenyans do of vVO2max) and then to v1500m in the same inter-
today. val training session (distance varying from 3000 to
In the sixties, Wasserman and McIlroy[21] invent- 200m).
ed the concept of anaerobic threshold (1964) as a Trainers used specific velocities from 800 to
pathological diagnostic tool. However, this concept 5000m to calibrate their interval training without
taking into account the physiological markers. Dan-
has not yet been used to delineate training velocity .
iels et al.[26] defined the parameter vVO2max as the
zones and training in this velocity range was done .
velocity associated with VO2max determined by an
as a form of Fartlek.
incremental work test on a treadmill. Furthermore,
In 1967, the Swedish physiologists Bengt Saltin .
. this vVO2max was found to be close to the average
and Per Olf Astrand[22] published data on VO2max
velocity sustained over 3000m.[26,27]
for athletes including the world record holder for
The concept of velocity associated with the
3000m (7 minutes 39 seconds), Kip Keino. They .
VO2max appeared at the beginning of the eighties
reported the highest value recorded for a runner (82 with Daniels et al.,[28] di Prampero[29] and the max-
ml/min/kg), not much higher than that reported in imal aerobic speed of the Montreal track test of
1937 by Robinson et al.[23] for the 2 mile world Lger and Boucher[30] (for review, see Billat and
record holder, Donald Lash who had a VO2max of Koralsztein[31]). This last test allowed one to have
. .
81 ml/min/kg. However, considering the vVO2max an estimate of VO2max (with an average energy cost
value reported (21.6 km/h) on the level treadmill, of running) and provided the minimal velocity elic-
these data were probably not overestimated. In 1955, iting VO2max in an incremental test.
Astrand published higher values[24] found in cross However, Karlsson et al.[32] had already shown
country skiers. However, despite these prestigious in 1967, that the speed exhausting runners in 4 min-

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
Aerobic Interval Training and Performance 19

Velocity on 5000 PR that they have available a very wide range of ve-
25 Velocity on 5000 NR
Velocity on 10 000 PR
locity and interval training from vMLSS to maximal
Velocity on 10 000 NR speed.
Another key factor, as we shall see in this re-
24 view, is strength and power development which
becomes more and more important in improving
Velocity (km/h)

performance over long distances (by decreasing the

23 cost of activity). This was emphasised by trainers
such as Percy Cerutty who trained Australian dis-
tance runners (Herb Elliott and John Landy) who
22 dominated the international competitions in the late
1950s and early 1960s. He asked these runners to
perform interval runs up sand-hills as well as under-
take extensive weight training sessions. He recom-
0 1 2 3 4 5 6 7 8 9 10 mended that all distance runners spend at least a
Distance (1000m) third of their training time in nonrunning activities,
in particular weight training, which can be organ-
Fig. 1. Variation in velocity during 5000 and 10 000m of the
previous record (PR) and of the new world record (NR) for the ised as an interval training (called circuit training).
5000 and 10 000m events. Cerutty emphasised that there were 2 important as-
pects during training: (i) to run at competitive speed
[not full distance]; and (ii) to train at high velocity
utes could be reduced to 80% (3 km/h), without continuously for the full distance. However, almost
decreasing their oxygen consumption. It is now everything was included in his programme.
well known that the minimal velocity which elicits
. This requirement of using weight training was
VO2max in a steady-state velocity condition is below confirmed recently by Paavolainen et al.[34] who
vVO2max (found in incremental 3-minute stages), reported that the velocity over 5km was positively
and is set just above the critical velocity, in the mid correlated with the maximal velocity, the contact
range between the velocity at the maximal lactate time and the stride rates over 20m (running start).
steady state (vMLSS) and vVO2max (the so-called v50 Both the velocities over 5 and 10km were correlated
velocity).[33] with the mean contact time of the constant velocity
Indeed, vVO2max is easy to measure with an in- laps during 5 and 10km. The ability of fast force pro-
cremental test and can be used as a reference for in- duction during maximal and submaximal running
terval training calibration. Moreover, out of compe- was related to both the 5 and 10km performance.[34]
tition season it seems better to refer to the velocities The same group of researchers also showed that
associated with particular physiological responses explosive strength training (various sprint, jump-
in the range from maximal lactate steady state to ing exercises, leg press and knee extensor-flexor
the absolute maximal velocity. The range of veloc- exercises) replacing 32% of the training volume
ities used in the race must nevertheless be taken induced a significant decrease in the 5km time.[35]
into consideration, since even world records are not This increase in performance was related to the im-
run at a constant pace. Figure 1 shows the velocity proved running economy and the velocity reached
change in the world record race over 10 000 and in an anaerobic treadmill running test.[36]
5000m, and one can appreciate that the competition Another new interval training technique could
is also a type of interval running. If we consider the be to use different sports (cycling and running for
training of the best middle- and long-distance run- instance) simultaneously in the same interval train-
ners of the 20th century (table III), it can be seen ing session. This has been used by triathletes to

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
20 Billat

Table III. Summary of some of the greatest champions training referenced with physiological marker velocitiesa
. .
Year; name; best performance vVO2max (km/h); vMLSS b Critical velocityc 90-100% >vVO2max d
. .
VO2max (ml/min/kg) vVO2max d
1920; Paavo Nurmi; 14 min 28 22.0; 75 15-20 km/day 20 100m, R = 4 400m at
sec over 5000m (20.7 km/h); 30 200m walked maximal velocity
min 6 sec over 10 000m (19.9 over 400m, R = 15
km/h); 1 training/day min rest
1950; Emil Zatopek; 13 min 23.5; 76.2 20 km/day 20 200m + 40 2 + 40 200m, r = 6 400m at 90% of
57.02 sec over 5000m (21.5 400m + 20 200m = 200m jogged maximal velocity
km/h); 28 min 54.02 sec over 10 200m trot; or 50 over 400m, R = 10
000m (20.8 km/h); 2 trainings/day 200m morning and min rest
1968; Kip Keino; 7min 39.05 sec 23.5; 80.0 5 45 min or 10 400m, r = 2 10 200m + 10
over 3000m (23.5 km/h); 13 min 6 60 min min jogged; or 6 100m + 4 80m at
36.05 sec over 5000m (22.1 800m, R = 3-5 90% of the maximal
km/h); 2-3 trainings/day min jogged velocities over the
distances, r = 300m
1972-1976; Lasse Viren; 13 min 24.0; 83.0 80km per 130km Fartlek (over 10 200m, r = 2 8 600m, r = 600m
16 sec over 5000m (22.6 km/h); week at 100 12-15km) min; or 6 walked
27 min 38 sec over 10 000m bpm 800m, r = 3-5
(21.7 km/h); 2-3 trainings/day min jogged
1984; Grete Waitz; 15 min 8 sec 21.0; 73.0 45 min-2h 20 min at CV in 60 min 6 1000m, r = 1 2 (10 300m), r =
over 5000m (19.8 km/h); 30 min every day of running (tempo min; or 5 100m walked, R = 5
59.08 sec over 10 000m (19.4 training) 1600m, r = 2 min walked
km/h); 2h 25 min 28 sec in min jogged; or 5
marathon (17.4 km/h); 2 2000m, r = 3
trainings/day min jogged
1986; Ingrid Kristiansen; 14 min 21.7; 76.0 45-150 min 2 15 min at CV in 90 5 1000m, r = 2 2 series of 5
37.03 sec over 5000m (20.5 every day min run (tempo training) min 100m, r = 200m
km/h); 30 min 13.07 sec over 10 walked, R = 400m
000m (19.9 km/h); 2h 21 min 6 walked
sec in marathon (17.9 km/h); 2
. . .
a vVO2max and VO2max were estimated from the runners personal best over 3000m (97% of vVO2max). CV was computed from their per-
sonal best over 3000 to 10 000m.
b Run over a distance (km/day or km/week) or for a duration (min) 80% vVO2max.
c Run over a distance (m or km) or for a duration (min) 85% vVO2max.
d Run over a distance (m or km) or for a duration (min).
bpm = beats per minute; CV = critical velocity; R = recovery between series (i.e. set of several repetitions); r = recovery between repetitions;
. .
vMLSS = velocity at the maximal lactate steady state; VO2max = maximal oxygen uptake; vVO2max = velocity at maximal oxygen uptake.

elicit longer VO2max in the same training session metabolic predictor for 100, 200 and 400m perfor-
and to become accustomed to working successive- mance, and that the peak oxygen uptake (VO2peak)
ly with different muscle masses. However, cross- was the best predictor for 800, 1500 and 5000m
training has been reported to be less effective in performance.[40] However, Spencer et al.[41] clearly
improving VO2max especially in highly trained in- demonstrated that even over 400m the role played
dividuals.[37,38] The ideal cross-training could be a by the aerobic energy system is more important. In-
mixture of aerobic and anaerobic interval training deed, VO2max was elicited in the last 20 seconds of
for both short-long sprint (400m run) and middle- a 400m run in 52 seconds (run at 170% of vVO2max).
distance runners (800 to 5000m).[39] The energy produced by oxidative phosphorylation
Multiple regression analyses have indicated that was 46% of the total energy versus 69 and 83% for
the maximal accumulated oxygen deficit was the best a 800m and 1500m run in 1 minute 58 seconds and

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
Aerobic Interval Training and Performance 21

4 minutes 2 seconds, respectively (regional perfor- of pVO2max) and all-out exercise (6 minutes) at 100%
mance). The maximal accumulated deficit was the of pVO2max. With active pause, it may be that, since
same over all these distances. This confirmed the the average power output is higher, intermittent ex-
hypothesis of Hill[42] explaining the decrease of ve- ercise is closer to continuous maximal exercise at
locity with time; the fact that the oxygen deficit 100% of pVO2max concerning muscle fibre type re-
was divided by a longer time decreased the power cruitment and enhancement of the glycogenolysis
output. As noted by Houmard et al.,[43] anaerobic since CP has less time to be reconstituted. More-
systems influence middle-distance performance in over, during intermittent exercise a lower glycogen
runners of similar abilities. depletion may be explained by a relative increase
Thus, all types of interval training should be in the contribution of lipids to oxidative metabo-
used, especially in middle-distance runners up to lism.[45] These could also increase levels of aden-
10km and perhaps, in the future, for marathon run- osine triphosphate (ATP), CP and citrate at the end
ners. The last lap of a 10 000m race is currently run of each rest period, which would suppress glycol-
in less than 1 minute (>24 km/h), well above the
. ysis in the early phase of the subsequent work pe-
runners vVO2max. riod.
As a consequence of the brief work intervals,
3. Aerobic Interval Training oxygen stored in myoglobin (400ml) can supply
half of the oxygen requirement. As underlined by
3.1 Immediate Responses to Aerobic
Christensen et al.,[9] there may be an increased avail-
Interval Training
ability of oxygen because of the reloading of myo-
Aerobic training is defined as an interval train- globin stores in the resting periods and subsequently
ing which elicits aerobic metabolism at a higher a greater aerobic energy output which would give
ratio than anaerobic metabolism. This can be esti- a higher ATP production per glucose unit compared
mated from the ratio between the accumulated ox- with lactate formation. In fact, for the purpose of
ygen deficit and the oxygen consumed in interval maximally taxing the oxygen-transport system, As-
training. trand and Rodahl[46] recommended intermittent 10-
second runs and 5-second pauses to reach VO2max.
3.1.1 Short Aerobic Interval Training .
Short aerobic interval training has been shown A well trained athlete (VO2max = 5.3 L/min) is able
to sustain this configuration of exercise for 30 min-
to prevent glycogen depletion by using lipids com- .
pared with continuous exercise performed at the utes with an effective run time at vVO2max of 20
same velocity. A high intensity exercise (100 to minutes (since the work : rest ratio was 1/2). The
. oxygen deficit (the difference between the amount
102% of pVO2max) performed continuously (time
limit = 4 to 6 minutes) or intermittently (112% of of oxygen needed and the amount of oxygen actu-
. ally consumed), was thought to be accounted for
pVO2max) for 60 minutes (15 seconds of work and
15 seconds of rest) did not deplete muscle fibres in by the utilisation of other energy stores such as
the same way.[44] The blood lactate level was equal high energy phosphates (e.g. phosphocreatine) and
to 2 mmol/L in the intermittent exercise versus 10 the oxygen bound to myoglobin.
mmol/L in the continuous one. Indeed, after 60 min- Olbrecht et al.[47] have demonstrated that in swim-
utes of intense interval training, a significant and ming, the velocities during brief intervals of 10-
similar depletion occurred in both type I and type second swims with rest periods of 10 seconds were
II (A + B) fibres. With continuous intense exercise higher than those corresponding to the same lactic
to exhaustion, glycogen depletion was more marked acid levels during continuous swimming, +11.2,
in type II (A + B) than type I fibres. In fact, inter- 4.2, 2.9 and 2.0% of the velocity at the 4 mmol/L
mittent exercise depletes muscle fibres in an inter- level on 50, 100, 200 and 400m, respectively. The
mediate way between continuous submaximal (50% velocity on the basis of the 4 mmol/L level was

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
22 Billat

obtained from the 2-speed test (2 400m).[47] This Gorostiaga et al.[51] showed that interval train-
short rest of 10 seconds allowed the swimmer to ing with repetitions of 30 seconds work at 100%
regenerate the myoglobin oxygen reserve but prob- vVO2max, separated by 30 seconds of rest, produced
ably not the CP reserve. With a longer rest (30 sec- a greater increase in VO2max than continuous train-
onds) the increase in the velocity compared with ing at 70% vVO2max. In this last study, both contin-
continuous work was 1.5-fold greater than those uous or intermittent training only elicited a VO2
obtained with a rest of 10 seconds. which was 70% of VO2max. All of these studies were
In rowing, short interval training with 15 sec- aimed at improving VO2max and were based on the
onds at the competition velocity (since the compe- assumption that the more specific the stimulus, (i.e.
tition lasts 6 minutes, it is probable that the rowers taxing the cardiovascular and aerobic enzymatic
. system to their maximum) the greater the improve-
elicit 100% of VO2max) and 15 seconds of rest has
also been tested.[48] This was performed as 5 sets ment. However, again, none of these studies mea-
of 5 repetitions. Each set was separated by 30 sec- sured the time the athlete spent at VO2max.
onds of rest. During the intermittent rowing, no sig- It is interesting to note that none of the previous
studies checked if their participants had reached
nificant differences were detected in any of the vari-
. .
ables measured between sets. Heart rate, VO2peak VO2max during the interval training session. How-
ever, Astrand et al.[8] reported that interval training
and blood lactate averaged 89, 78 and 32%, respec- .
of 2 minutes run at vVO2max alternated with inac-
tively, of values measured during the continuous .
tive rest of the same duration elicited a VO2 equal
incremental exercise test. Therefore, with rowing, .
to 95% of VO2max accompanied by a very low blood
the investigated 15/15 intermittent exercise model
lactate level (2.2 mmol/L). The same study also
demands relatively high aerobic loading and low
reported that interval training using shorter repeti-
glycolytic activity. Gullstrand[48] concluded that this .
tions (15 seconds at vVO2max alternated with 15
type of interval training may be considered as an .
seconds of complete rest) did not bring the VO2 to
alternative model for training which would allow
maximum levels.
the rowers to work for prolonged periods of time
Billat et al.[52] reported that in a 3030-second,
at values slightly above competition intensity.
short interval training, the 30 seconds of recovery is
The short interval training currently being used . .
active (50% vVO2max), and runners stayed at VO2max
by runners and studied by researchers is the 3030- even during the recovery period from the fifth to the
second work : rest pattern. Thirty years ago, Edwards .
last (18th) repetition of a 30-second run at vVO2max,
et al.[49] had already measured, breath by breath, .
and a 30-second run at 50% of vVO2max. This short
oxygen kinetics in 30-second interval training dur- interval training with active pauses allows individ-
ing active pause performed at 50% of a work rate .
uals to sustain VO2max for 10 minutes (83% of total
associated with a time limit of 6 minutes (close to .
. . time run at vVO2max). The average blood lactate
the power output at VO2max: pVO2max). This was end value was 7.4 1.8 mmol/L. Runners reached
carried out on a cycle-ergometer. The work : pause .
VO2max during the intermittent exercise, the associ-
ratio was 1 and the interval duration was 30 sec- ated blood lactate was at a steady-state level and from
onds. They emphasised that VO2 (ventilation and the third to the sixth minute was below 4 mmol/L.
heart rate) remained high throughout the 20-second Hence, for at least 1 minute, these 5 runners were
recovery intervals and did not fall appreciably until at VO2max with only 4 mmol/L of blood lactate. This
30 seconds after the end of the last work period. is interesting since in previous studies which have
With passive rest in a 3030-second interval train- examined blood lactate accumulation during inter-
. .
ing performed at 120% vVO2max, VO2 reached only mittent exercise, it has been reported that only a
. .
70% of VO2max during the work periods (and 14 high value of blood lactate accompanies a VO2 at
mmol/L of blood lactate).[50] its maximum value.[8] This is because of the fact

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
Aerobic Interval Training and Performance 23

that these studies have used long 2- to 3-minute at vVO2max (intervals set at 50% of the individuals
. .
intervals to elicit VO2max with complete rest be- time to exhaustion at vVO2max) separated by active
tween repetitions. Therefore, when using the 30- pauses of the same duration run at 50% vVO2max.
second rest/exercise intervals with inactive pause Saltin et al.[4] found similar blood lactate values at
they did not reach VO2max. the end of 30 minutes of intermittent cycling with
Interval training performed at velocities close to intervals of 30 seconds of supramaximal work
. . .
the velocity associated with VO2max (vVO2max) may (400W, %VO2max not specified) and 60 seconds of
. rest. By using the same ratio (1 : 2) between rest
maximise the improvement in VO2max, as well as
result in significant improvements in mitochondrial and exercise but by increasing the time by 2 (i.e.
density.[53] In fact, in addition to these aerobic (O2 60 seconds work and 120 seconds rest), their par-
transport) training benefits, interval training stim- ticipants tripled their end blood lactate values (18
ulates the rate of lactate removal which depends mmol/L).
However, if we consider, endurance (time to ex-
directly on its level (i.e. the greater the level, the .
greater the removal).[53] Therefore, interval train- haustion) at VO2max, it would be interesting to com-
pare the influence of training using protocols which
ing that increases blood lactate levels will also stim- .
elicit VO2max but which are run at different veloc-
ulate lactate removal. For this reason Brooks et al.[53] .
recommended activity during the rest interval to ities (v50 versus vVO2max for instance). In ac-
stimulate this lactate removal and hence avoid blood cordance with Noakes,[19] the benefits of training
lactate accumulation. Indeed, in 1937 Newman et also depend on the distance covered at a high ve-
al.[54] had already noticed that the removal of lac- locity determining the muscular adaptation max-
tate, accumulated after exhausting exercise, was en- imising the number of powerful muscle contrac-
tions. For this purpose, the intermittent exercise
hanced if the individual continued to exercise dur- .
training at vVO2max, not only allows the cardiovas-
ing recovery, but at lower intensity, which normally
cular function to be stimulated at its maximum (at
did not accumulate lactate. This information, con- .
VO2max) for a longer time, but allows the run to be
firmed in 1975 by Belcastro and Bonen,[55] was ap-
made at a higher velocity (+1.6 km/h). Therefore,
plied in training programmes for elite athletes in
both from the cardiovascular and muscular adapta-
the 1950s. Despite high lactate production at the .
tion point of view, intermittent exercise at vVO2max
high velocities used in interval training (i.e. above
is likely to produce increased performance for mid-
the lactate threshold) walking or jogging in the rest
dle-distance runners.
phase of intermittent exercise would tend to stim- .
Before speculating on the cause of VO2max im-
ulate oxidative recovery.[56] Therefore, we suggest
provement from a given training design, it is essen-
that active, rather than passive, pauses between the
tial to examine the effect of this stimulus on car-
intervals of hard work will not only elicit and main-
. diovascular and metabolic responses. In the absence
tain VO2max, but will also stimulate lactate removal of this information, we can only hypothesise that
whilst remaining close to the maximal blood lac- the benefit of these training procedures on aerobic
tate steady state. .
capacity (and especially on VO2max) is dependent
The blood lactate values obtained at the end of .
not only on the time spent at VO2max but also on
the intermittent tests (6.8 2.2 mmol/L) are of a the distance run at a high velocity. With this in
similar magnitude to those reported by Gimenez et mind we are then able to discriminate between the
al.[57] These authors used a 45-minute exhaustive benefits gained from either interval or constant load
square wave endurance exercise test composed of
. tests.
9 repetitions with 1 minute at vVO2max and 4 min-
utes at 50% vVO2max. This blood lactate level is 3.1.2 Long Aerobic Interval Training
also in accordance with that reported by Billat et In addition to the above studies, it has been stated
al.[13] during running with interval training periods for a long time now that a typical endurance train-

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
24 Billat

ing programme consisting of repeated 1- to 8-minute ning events.[7,25,27,31,61,62] Optimal improvement in

runs at 90 to 100% vVO2max is the most effective cardiorespiratory fitness is thought to be induced
programme for improving VO2max and performance by training at an intensity corresponding to 90 to
for middle-distance runners.[58] During interval train- 100% of VO2max.[63]
ing lasting 7 2 minutes at pVO2max (with recovery The duration of intermittent effort varies con-
periods long enough to allow heart rate to return to siderably depending upon the author. In training
. .
130 bpm), VO2 reached 70% of VO2max in the first for track and field events, intervals between 10 sec-
minute and 100% of VO2max in the second minute. onds to 3 minutes, generally spaced by inactive rest,
Therefore, with this active recovery procedure (heart have been investigated but have not been referenced
rate being at 130 bpm at the end of the recovery), to individual possibilities to sustain high intensity
. .
it took only 1 minute to reach VO2max when running exercise (i.e. time to exhaustion at vVO2max).[31]
. .
at 100% vVO2max. The total time spent at VO2max Training with intermittent runs at 60 and 100% of
was 7 minutes in all 7 work intervals. However, the vVO2max (with a duration equal to half of the indi-
. .
time spent at VO2max in this 2-minute interval train- vidual time to exhaustion at vVO2max) allowed long-
ing at vVO2max was no longer than that during an distance runners to double the distance covered at
all-out run at a submaximal velocity inducing a slow vVO2max compared with continuous training runs
. .
phase of oxygen kinetics (90% vVO2max).[59] at vVO2max.[13] More recently, Billat et al.[64] and
Therefore, 30 seconds appears to be the longest Smith et al.[65] have reported that only 1 session per
interval duration allowing work at pVO2max to elicit week (for 4 weeks) of this kind of individualised
VO2max even in the recovery period. One or 2 min- interval training (50 to 75% of the time to exhaus-
. . .
utes at pVO2max induces high blood lactate levels tion vVO2max) significantly increased vVO2max in a
because of the depletion of CP and the use of the group of middle- and long-distance runners.
oxygen myoglobin-bound oxygen reserve. A recov- To calibrate long interval training, time to ex-
ery period of the same duration as the work period haustion at the velocity associated with VO2max could
(1 to 2 minutes) allows the rephosphorylation of be a new parameter which could be used to deter-
CP but decreases oxygen consumption. Moreover, mine a rational basis for interval training in elite
passive or a low work rate (40% of VO2max) in the middle- and long- distance runners.[13] Therefore,
recovery phase, allows the CP store to be replaced the use of time to exhaustion (time limit = tlim) at
and avoids a high blood lactate level. However, vVO2max could allow elite long-distance runners to
using a higher work rate during recovery, Fox et run longer distances at vVO2max during interval train-
al.[60] demonstrated that when CP renewal is par- ing. Since time to exhaustion at vVO2max has been
tially blocked by performing aerobic work (60% previously reported to be very different among run-
. .
vVO2max) rather than resting during the relief inter- ners with the same vVO2max, we hypothesised that
vals, a greater proportion of the energy needed dur- this could be a rational basis for determining the
ing the work intervals was supplied by the anaero- length of the work intervals. We have compared the
bic lactic metabolism. distances run with the physiological responses at
Hence, long interval training is difficult to man- the end of 2 interval training sessions that were
age if one wants to avoid acidosis. To calibrate the performed on a treadmill and run at vVO2max. Two
intensity of interval training, coaches often refer to exhaustive intermittent tests were performed, one
the running velocity associated with the achieve- using a standard 2-minute duration for the alternat-
ment of VO2max during an incremental treadmill ing exercise and recovery periods, the other using
test (vVO2max) and to the running velocity at the durations that were individually determined based
onset of blood lactate accumulation (vOBLA). These on time spent at vVO2max. The study involved 16
. .
have both been reported to be relevant indicators male good level runners (VO2max and vVO2max
of performance for middle- and long-distance run- were 69.1 4.3 ml/min/kg and 21.4 1 km/h, re-

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
Aerobic Interval Training and Performance 25

. .
spectively).[13] When the intermittent exercise train- over time (VO2 reached = 69.5 5.0 vs VO2max of
ing stimulus was standardised by alternating dura- 74.9 3.0 ml/kg/min). In other words, highly trained
. .
tion equal to 50% of the time spent at vVO2max long-distance runners did not exhibit the VO2 slow
(with equal periods of recovery run at 60% vVO2max), component when performing exhaustive, supra-
. .
the total time at vVO2max was 2.5 times the contin- critical velocity runs at 90% vVO2max, significant-
uous time limit whatever the value of time to ex- ly above their critical and lactate threshold velocities
. .
haustion at vVO2max. This means that all of the run- (at 82 and 86% of vVO2max, respectively). Instead,
ners were able to run 5 repetitions at 50% of their these runners maintained a steady-state VO2 below
. .
continuous time to exhaustion at VO2max.[13] VO2max, such that the time to exhaustion at 90% of
We are going to consider now, the very long vVO2max for these runners was positively corre-
interval training at velocities between the vMLSS lated with the critical velocity expressed as a per-
. .
and vVO2max. Continuous running above the run- centage of vVO2max. Critical velocity is known to
ning velocity at which the critical velocity is at- correspond to an exercise intensity that lies between
tained[66] could be more efficient to elicit VO2max of the work rate associated with the lactate threshold
. .
longer duration than interval training at vVO2max.[67] and VO2max.[33] It has been suggested that this in-
Indeed, previous studies reported that during se- tensity is comparable to that achieved in a com-
vere exercise an additional slow phase of VO2max petitive 10km race.[68,71] In addition, Gaesser and
(the VO2 slow component) is superimposed upon Poole[67] have proposed that during prolonged ex-
. . .
the underlying VO2 kinetics and VO2 continues to ercise at intensities above the critical velocity, VO2
increase until the end of the test or until exhaustion, would continue to rise until VO2max is reached.
. .
and will possibly drive VO2 to the VO2max.[68,69] However, the results from the study of Billat et
. .
Therefore, by using this VO2 slow component al.[70] reported that a VO2 slow component was not
phenomenon, it might be possible to elicit VO2max expressed by high level long-distance runners dur-
for a longer time, provided the individuals run ing exhaustive supra-critical velocity runs at 90%
for a sufficiently long period at this supra-critical of vVO2max. Indeed, although these runners were
velocity.[33] In this last study, the work rate chosen assigned to run at a work rate which was 5% above
was in the mid range of the work rates associated their critical velocity (90% of vVO2max), they reached
. . .
with maximal lactate steady state and vVO2max a VO2 steady state at an average of 93% of VO2max
(called v50). In fact, 5 runners out of 8 reached for 17 minutes and did not demonstrate a progres-
. .
VO2max during this severe constant load run. There- sive increase in VO2 over time. At the end of this
fore, continuous running at v50 can be used to all-out run they had a blood lactate level of 6.5
elicit VO2max in a group of middle level runners mmol/L, a value of the same magnitude as those
(15 minutes 30 seconds for 5000m) having a high obtained in the present study.
. .
fractional VO2max at the lactate threshold but not a Indeed, the VO2 slow component appears in in-
. .
very high VO2max. Indeed, most of this group of dividuals having a smaller VO2max rather than elite
runners (6 of 8) developed a VO2 slow component athletes. Therefore, it is possible that continuous
during track running at v50. However, this was severe exercise (i.e. v50) cannot be used to stim-
not the case in a study of high level runners having ulate VO2max for highly trained runners who al-
a similar endurance capacity (i.e. vMLSS at 84% of ready have a high VO2max (>70 ml/min/kg).
. .
vVO2max) but with a VO2max which was 23% greater Most previous investigations have described a
(75 vs 61 ml/min/kg).[70] slow phase of VO2 during intense cycling exercise
Moreover, Billat et al.[70] reported that in a supra- performed by untrained individuals.[68,72-74] For un-
critical velocity run (90% vVO2max),[66] these 14 trained individuals, endurance training has been
highly trained long-distance runners reached a VO2 shown to reduce the magnitude of the slow com-
steady state, but did not reach their VO2max levels ponent,[75] and it can be suggested that for them,

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
26 Billat

long interval training at v50 can be used to stimulate critical velocity or at a mid-way between vOBLA
. .
VO2max. This kind of continuous supra-critical ve- and vVO2max) can be used as an intensity to elicit
locity is known as tempo training and is often used VO2max for long periods of time. Moreover, the in-
for training long-distance runners. However, few dividualisation of the duration of the repetition
trainers are aware of the VO2 slow component phe- avoids early blood lactate accumulation. The aver-
nomena. Continuous work at v50 allows less time age duration was, on average, 5 minutes at 92%
. .
to be spent specifically at VO2max, much less than vVO2max with a recovery of 2:30 at 50% vVO2max.
during the 30/30 seconds light/heavy exercise in- However, this is probably not sufficient to elicit
. .
tervals (3 minutes of time spent at VO2max for v50 VO2max in high level runners where shorter and
versus 10 minutes in 18 5 repetitions of 30/30 higher velocity interval training may be prefera-
seconds). Moreover, the blood lactate response was ble.[19,76]
more pronounced in the 50 run compared with the 3.1.3 Interval Training to Estimate Performance
interval training (6.8 2.2 vs 7.5 2.1 mmol/L, Interval training can be used to estimate perfor-
respectively). mance. In 1997, Babineau and Lger[77] reported
However, v50 can be used for an intermittent that performance over 5000m was well predicted
protocol rather than a continuous one.[59] Demarie by the maximal cruising speed on a 6 800m run
et al.[59] have demonstrated that lower level runners (30 seconds rest). Maximal cruising speed was the
. .
(VO2max = 60 ml/min/kg) reached VO2max at 92 highest average speed that could be maintained over
2% of vVO2max, a velocity between the onset of all of the work intervals in a single training session.
blood lactate accumulation and vVO2max (v50%). For both runners and multi-sports participants, 6
The long interval training was set at half the time 800m interval training was more closely corre-
to exhaustion at v50% as described previously by lated with performance [the recovery between rep-
Billat et al.[13] for interval training at vVO2max. The etitions (r) = 0.95, p < 0.001, n = 23] than 12
recovery was active (50% vVO2max) and the work 400m (15 seconds rest) [r = 0.90, p < 0.001, n = 23]
: pause ratio was 2/1. A time to exhaustion of 10:23 and 3 x 1600m (60 seconds rest) [r = 0.93, p <
1:26 min:sec ranging from 8 to 13 minutes for 0.001, n = 23]. Total distance run was always 4800m
the continuous run at v50% was found. Hence, the and the work : pause ratio was equal to 5. For run-
exercise periods of the intermittent exercise ranged ners only, the cruising speed over 1600m corre-
from 4:00 to 6:30 min:sec and the recovery periods sponded to the specific velocity during the 5000m
from 2:00 to 3:15 min:sec. The sum of the exercise run. This study was adapted to trainers methods
periods resulted in an average time limit of 19:38 which uses the target specific velocity of the com-
5:10 min:sec for the interval training. Runners petition (those of the last season to target those for
performed 3 1 repetitions of these long interval the present season) as a reference.
. 3.2 Long Term Effects of Aerobic
All participants showed similar VO2 kinetics,
. Interval Training
with maximal VO2 throughout all exercise periods.
VO2max was sustained for twice as long in this long 3.2.1 Long Term Effects of Short Aerobic
interval training than in the continuous run at the Interval Training
same velocity (v50%) [10:23 5:51 vs 5:07 In some cases short interval training has been
3:03 min:sec for the intermittent versus continuous considered to be less effective in enhancing VO2max.
run, respectively]. Blood lactate accumulation was The effectiveness of the different types of interval
less in the intermittent (6.5 2.2 mmol/L) versus training has been compared: short (30 to 40 repeti-
the continuous run (7.8 2.2 mmol/L). The authors tions of 15 seconds work, 15 seconds rest) versus
concluded that intermittent training at a velocity long (4 to 6, 4 minutes work, 2 minutes rest) per-
. .
which elicits a VO2 slow component (above the formed at 130 and 115% of vVO2max, respectively,

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
Aerobic Interval Training and Performance 27

for recreational runners (54.8 3.0 ml O2/kg/min).[78] frequency and/or respiratory exchange ratio during
The third kind of training was continuous distance submaximal run were unaltered with training.[78]
running at 90% vVO2max for 20 to 30 minutes. The In 1971, Davies and Knibbs[79] insisted that to ef-
total duration of the different types of training fect an improvement in VO2max an individual must
(performed by 3 paired groups, 3 days per week be prepared to work at or close to their VO2max for
over a 6 week period was similar. The VO2max im- prolonged periods of time, and even then improve-
provement was significantly higher for the long in- ment might be disappointingly small. Recently, in
terval training and the continuous running (+6%) his training book for running, Daniels et al.[28] shared
versus the short interval training (+3.6%). This this point of view. This is especially important for
could be because of the fact that complete rests previously trained runners compared with untrained
. . .
were used that prevented VO2 reaching VO2max in individuals; even for untrained individuals, VO2max
the short interval training. Moreover, the continu- was stable after 4 weeks of training and did not
ous running was performed at a high velocity (90% increase until the eighth week (end of training).[28]
. .
of vVO2max making them reach 92.5% VO2max) to Interval training was more effective for increas-
. .
induce a VO2 slow component and increase VO2 ing rates of fatty acid oxidation than continuous
towards VO2max.[59] In addition to this intense pro- training, despite lower total energy expenditure. With
gramme, runners had slow-pace runs at 78% of max- continuous training, the relative increase in rates
imal heart rate (HRmax). The greatest improvement of respiration with pyruvate and palmityl-carnitine
was in the time limit at 85% of the pretraining (80% were equal, implying that the activity of enzymes
post-training). The largest increase was seen in the involved exclusively in pyruvate oxidation (i.e. py-
continuous run group where time to exhaustion in- ruvate dehydrogenase) increased in proportion to
creased by 94% from 35 to 68 minutes; for long those involved in fatty acid oxidation (i.e. enzymes
interval training time increased by 67% and for of -oxidation). In contrast, with interval training,
short interval training time to exhaustion increased the relative increase in the rate of respiration was
by 65%. The continuous run seems to be more ef- greater with palmityl-carnitine, compared with py-
fective in enhancing both VO 2max and endurance ruvate, implying that activity of enzymes involved
at a submaximal velocity; however, this probably in fatty acid oxidation increased to a greater extent
comes from the inactive rest used in the interval than those involved in pyruvate oxidation. Indeed,
training. Submaximal VO2 at the same velocity (85% adaptations of the -oxidation pathway are most
VO2max of pretraining value) was decreased in re- likely induced with exercise that promotes the use
lation to ventilation. Blood lactate measured at the of fats and a high flux rate for fatty acids. Interval
end of the submaximal run was decreased from 6.6 training may promote these conditions.[80] This con-
0.4 to 5.1 0.3 mmol/L after training. firms the hypothesis proposed 20 years earlier by
This study clearly showed that moderately train- Essen[44] and Essen et al.[45]
ed recreational runners can improve both running With repeated high intensity bouts (close to
. .
economy and VO2max within a relatively short pe- pVO2max), lactate and citrate may inhibit glycoge-
riod (6 weeks) by exchanging parts of their con- nolysis during later bouts, resulting in an increased
ventional aerobic distance training with more in- reliance on fatty acid oxidation.[45] Essen et al.[45]
tensive distance or long interval training. A reduced have used 60 minutes of intermittent intense exer-
pulmonary ventilation following training correlated cise, 15 seconds work at vVO2max and 15 seconds
significantly with the improved running economy, rest, compared with 60 minutes of continuous ex-
suggesting that ventilatory adaptation may contrib- ercise at a load (mean 157W) i.e. 50% of VO2max
ute to the improved running performance. Other selected to yield the same integrated VO2. Thus,
potential factors such as percentage of type I fibres the same total amount of work was performed dur-
in the vastus lateralis muscle, stride length, stride ing both types of exercise. She demonstrated that

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
28 Billat

at similar high workloads less glycogen is utilised spent (post-training) at vVO2max. Good level run-
and lipids contribute more to oxidative metabolism ners increased their vVO2max from 20.5 0.7 to
when exercise is performed intermittently (15 sec- 21.1 0.8 km/h (p = 0.02) after 4 weeks of 1 inter-
onds work, 15 seconds rest) than continuously. The val training sessions at vVO2max per week consist-
overall metabolic response to intermittent exercise ing of 5 50% of tlim at vVO2max with a work : active
is more similar to continuous exercise at about half recovery ratio of 1/1. Recovery was run at 50%
the load than at an equally high workload. There- vVO2max. The second intensive training was per-
fore, high intensity interval training resulted in a formed at the maximal blood lactate steady state: 2
smaller depletion of glycogen and a larger deple- 20 minutes with a work : active recovery ratio of
tion of intramuscular triglycerides compared with .
4/1, that is, 5 minutes at 60% vVO2max.[65]
low intensity continuous stimulation. With the in- .
The duration of interval training at vVO 2max
terval training a repeated stimulation of fatty acid can be longer: 60 and 75% of the time limit at
oxidation might have led to an up-regulation of this .
vVO 2max. [64] Using this interval training protocol
pathway, resulting in a greater stimulation of mito- 2 times per week, the performance over 3000m,
chondrial respiration in the presence of fatty acids.[80] . .
vVO 2max, VO 2max and time limit at the previous
Henriksson and Reitman[80] showed that inter- .
. vVO2max, increased after only 4 weeks.[64] As in the
val training (5 4 minutes at 10% VO2max with a study of Billat et al.,[64] Smith et al.[65] used a work
rest of 2 minutes in between) enhances the oxida- : rate ratio of 1 and recovery was run at 60% of
tive capacity of type II fibres compared with a con- .
vVO2max. The third training session was a recovery
tinuous exercise of the same duration performed at .
. session: 30 minutes at 60% of vVO2max. In good
79% of VO2max. These data are interesting for mid- level athletes it may not be necessary to use such
dle-distance runners who have more type II fibres long interval durations since in less than 2 min-
with a high oxidative capacity. However, only con- .
. utes,[33] oxygen reaches VO2max when they run at
tinuous training improved the whole body VO2max . .
vVO2max, and the time limit at vVO2max is higher
(+12%, p < 0.01).
than 4 minutes.[83] Smith et al.[65] tested runners
Tabata et al.[81] reported that supramaximal ex-
. with performances of 10 minutes over 3000m, i.e.
ercise (8 20 seconds at 170% VO2max with a 10-
. a middle level runner. It has been reported that ox-
second rest) enhances VO2max after 7 weeks where
. ygen kinetics is accelerated with training.[73,84,85]
individuals (VO2max 53 ml/min/kg) carried out 5
Using an individualisation of the duration of inter-
training sessions in 5 days. This very short interval
val training with reference to the time limit at a
training (less than 5 minutes of effective work) al-
given velocity allows the athlete to run the same
lowed them to increase both maximal accumulated
number of repetitions (5) despite the great inter-
oxygen deficit, an indicator of the anaerobic capac-
. variability of the time limit of the continuous exer-
ity (+28%) and VO2max (+13%).[81] The very short
pause means that the average power output is still
. . Importantly, the performance improvement in
very high (115% VO2max) and allows VO2 to in-
. all individuals in the study by Smith et al.,[65] was
crease to VO2max as demonstrated 1 year later by
uniform despite the heterogeneity (high coefficient
the same authors.[82] .
of variation) of vVO2max (18 to 22.7 km/h) and per-
3.2.2 Long Term Effects of Long Aerobic formance over 3000m (9 to 11 minutes). Therefore,
Interval Training for coaches with a group of runners with heteroge-
As seen in section 3.1, 2 longitudinal studies[64,65] neous levels, it may be useful to apply the time
calibrated the long interval training with reference limit to calibrate interval training work duration.
to the time limit at vVO2max and have reported a Burke et al.[86] demonstrated that if the exercise
. .
rapid increase of vVO2max (4 week, 2 interval train- intensity is the same (95% VO2max), interval train-
ing sessions per week) with no decrease in the time ing enhances VO2max (+7%) and blood lactate ve-

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
Aerobic Interval Training and Performance 29

locity (+25%). These changes appear to be inde- in ventilation. As hypothesised by Dempsey,[93]

pendent of the length of the interval (30 seconds or this suggests that there is little adaptability in the
2 minutes with a work : pause of 1). However, these pulmonary system to physical training for several
results were obtained in 21 female physical educa- months even if the enhancement of maximal aero-
tion students having only 40 and 43 ml/min/kg, bic power is huge.
respectively before and after the training period of
7 weeks at 4 times a week. The blood lactate thresh- 4. Conclusion
old velocity increased from 64 to 78% VO2max.
. In conclusion, as stated by Astrand and Rodahl[46]
It may be reasonable to assume that the high VO2 it is an important but unsolved question which
obtained during some forms of intermittent training type of training is most effective: to maintain a
leads to a significant stress on the aerobic system and .
. level representing 90% of the VO2max for 40 min-
results in the large increase in VO2max.[46,52,82] How- .
utes, or to tax 100% of the VO2 capacity for about
ever, it may be possible, as demonstrated by Billat
. 16 minutes. Today, this is still an open question.
et al.,[64] that vVO2max increases because of the de-
. Before beginning longitudinal studies to try to an-
crease of running economy and not of VO2max, es- swer this question, it is important to determine the
pecially with athletes who already have a good level
. metabolic response solicited by the different inter-
of VO2max (above 70 ml/min/kg). Even if interval
. . val training protocols used by trainers.[26]
training taxes VO2max, the increase in VO2max is not
. Even if optimal improvement in cardiorespira-
certain. To have some chance of increasing vVO2max,
tory fitness is thought to occur from training at an
in this case, the distance run at a high velocity may .
. intensity corresponding to 90 to 100% of VO2max,[46]
be a determinant for vVO2max and performance im-
. this central factor of performance is not the only
provement,[87] since vVO2max is related more to per-
. one to induce its improvement. Consequently, time
formance than VO2max.[7] However, as previously .
spent at VO2max is not the only parameter to be
demonstrated by Foxs group[88,89] almost 30 years
taken into account to judge the efficiency of a cer-
ago, it is more important to obtain improvement in
. tain pattern of interval training on the improvement
VO2max intensity than distance.[90] .
of VO2max and performance.

3.2.3 Hypoxaemia Induced by Exercise and Acknowledgements

Interval Training
Maximal interval training has been shown to This study was supported by grants from la Caisse Cen-
contribute to the development of arterial oxygen trale des Activits Sociales dElectricit et Gaz de France.
desaturation (hypoxaemia)[91] and less hyperven-
tilation during heavy exercise.[92] However, some References
. 1. Wells CL, Pate RR. Training for performance of prolonged
athletes with high VO2max do not develop arterial de- exercise. In: Lamb DR, Murray R, editors. Perspectives in
saturation during heavy exercise.[93] Interval training exercise science and sports medicine. Indianapolis (IN):
Benchmark Press, 1988: 357-91
was performed for 12 weeks (4 days/week) on a cycle
. 2. Reindell H, Roskamm H. Ein Beitrag zu den physiologischen
Grundlagen des Intervall training unter besonderer Berck-
ergometer at 100% VO2max for 3 minutes with 2 min-
. . sichtigung des Kreilaufes. Schweiz Z Sportmed 1959; 7: 1-8
utes of active recovery (50% VO2max). VO2max was 3. Reindell H, Roskamm H, Gerschler W. Das Intervall training.
increased from 50.9 5.6 to 61.6 ml/min/kg (+19%) Munchen (Germany): John Ambrosius Barth Publishing, 1962
4. Saltin B, Essen B, Pedersen PK. Intermittent exercise: its phys-
whereas maximum expiratory volume (VEmax) in- iology and some practical applications. In: Joekle E, Anand
creased only slightly during the early weeks of train- RL, Stoboy H, editors. Advances in exercise physiology. Med-
ing. Moreover, interval training also induced lower icine sport series. Basel: Karger Publishers, 1976: 23-51
5. Hill AV. Muscular movement in man. New York (NY): McGraw-
alveolar partial pressure (PAO2) and less hyperven- Hill, 1927
tilation during heavy exercise. Half of the reduc- 6. Ettema JH. Limits of human performance and energy produc-
tion. Int Z Angew Physiol 1966; 22: 45-54
tion in arterial oxygen saturation (SaO2) with in- 7. Daniels J, Scardina N. Interval training and performance.
terval training can be accounted for by the variation Sports Med 1984; 1: 327-34

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
30 Billat

8. Astrand I, Astrand PO, Christensen EH, et al. Intermittent mus- 31. Billat V, Koralsztein JP. Significance of the velocity at V O2max
cular work. Acta Physiol Scand 1960; 48: 448-53 and its time to exhaustion at this velocity. Sports Med 1996;
9. Christensen EH, Hedman R, Saltin B. Intermittent and contin- 22: 90-108
uous running. Acta Physiol Scand 1960; 50: 269-86 32. Karlsson J, Astrand PO, Ekblom B. Training of the oxygen
10. Astrand I, Astrand PO, Christensen EH, et al. Circulatory and transport system in man. J Appl Physiol 1967; 22: 1061-67
respiratory adaptations to severe muscular work. Acta Physiol 33. Billat V, Blondel N, Berthoin N. Determination of the velocity
Scand 1960; 50: 254-8 associated with the longest time to exhaustion at maximal
11. Astrand I, Astrand PO, Christensen EH, et al. Myohemoglobin oxygen uptake. Eur J Appl Physiol 1999; 80: 159-61
as an oxygen-store in man. Acta Physiol Scand 1960; 48: 34. Paavolainen LM, Nummela AT, Rusko HK. Neuromuscular
454-60 characteristics and muscle power as determinant of 5-km run-
12. Medbo JI, Tabata I. Relative importance of aerobic and anaer- ning performance. Med Sci Sports Exerc 1999; 31: 124-30
obic energy release during short-lasting exhausting bicycle 35. Paavolainen L, Hakkinen K, Hamalainen I, et al. Explosive-
exercise. J Appl Physiol 1990; 67: 1881-6 strength training improves 5-km running time by improving
13. Billat V, Pinoteau J, Petit B, et al. Calibration de la dure des running economy and muscle power. J Appl Physiol 1999; 86:
rptitions . dune sance dinterval training la vitesse 1527-33
associe V O2max en rfrence au temps limite continu. Sci 36. Rusko HK, Nummela A, Mero A. A new method for the evalu-
Motricit 1996; 28: 13-20 ation of anaerobic running power in athletes. Eur J Appl Phys-
14. Astrand PO, Saltin B. Oxygen uptake during the first minutes iol 1993; 66: 97-101
of heavy muscular exercise. J Appl Physiol 1961; 16: 971-6 37. Tanaka H. Effects of cross-training. Sports Med 1985; 18: 330-9
15. Karlsson J, Saltin B. Oxygen deficit and muscle metabolites in 38. Foster C, Hector LL, Welsh R, et al. Effects of specific versus
intermittent exercise. Acta Physiol Scand 1971; 82: 115-22 cross-training on running performance. Eur J Appl Physiol
16. Hermansen L. Anaerobic energy release. Med Sci Sports Exerc 1995; 70: 367-72
1969; 1: 32-8 39. Tanaka H, Swensen T. Impact of resistance training on endur-
17. Fox EL, Billings CE, Bason R, et al. Improvement of physical ance performance. Sports Med 1985; 6: 266-70
fitness by interval training, II: required training frequency. 40. Weyand PG, Cureton KJ, Conley DS, et al. Peak oxygen deficit
USA: Medical Research and Development Command, Office predicts sprint and middle-distance track performance. Med
of the Surgeon General, US Army; 1967 Apr. Report no.: RF Sci Sports Exer 1994; 9: 1174-80
2002-3 41. Spencer MR, Gastin PB, Payne WR. Energy system contribution
18. Mathews DK, Fox EL, Bartels RL, et al. Improvement of phys- during 400 to 1500 metres. News Stud Athlet 1996; 4: 59-65
ical fitness by interval training, I: relative effectiveness of 42. Billat V, Renoux JC, Pinoteau J, et al. Times to exhaustion at
short and long distance running. USA: Medical Research and 100% of velocity at V O2max and modelling of the time-
Development Command, Office of the Surgeon General, US limit/velocity relationship in elite long-distance runners. Eur
Army; 1966 Nov. Report no.: RF 2002-2 J Appl Physiol 1994; 69: 271-3
19. Noakes T. Lore of Running. Champaign (IL): Leisure Press, 43. Houmard JA, Costill DL, Mitchell JB, et al. The role of anaer-
1991 obic ability in middle distance running performance. Eur J
20. Daniels J. Jacks Daniels formula. Champaign (IL): Human Ki- Appl Physiol 1991; 62 (1): 40-3
netics Publishers, 1998 44. Essen B. Glycogen depletion of different fibre types in human
21. Wasserman K, McIlroy MB. Detecting the threshold of anaer- skeletal muscle during intermittent and continuous exercise.
obic metabolism in cardiac patients during exercise. Am J Acta Physiol Scand 1978; 103: 446-55
Cardiol 1964; 14: 844-52 45. Essen B, Hagenfeldt L, Kaijser L. Utilization of blood-born and
22. Saltin B, Astrand PO. Maximal oxygen uptakes in athletes. J intra-muscular substrates during continuous and intermittent
Appl Physiol 1967; 23: 353-8 exercise in man. J Physiol (Lond) 1977; 265: 489-506
23. Robinson S, Edwards HT, Dill DB. New records in human 46. Astrand PO, Rodahl K. Textbook of work physiology. 3nd ed.
power. Science 1937; 85: 409-10 New York (NY): McGraw-Hill, 1986
24. Astrand PO. New records in human power. Nature 1955; 176: 47. Olbrecht J, Madsen O, Liesen H, et al. Relationship between
922-3 swimming velocity and lactic concentration during continu-
25. Billat V. Use of blood lactate measurements for prediction of ous and intermittent training exercise. Int J Sports Med 1985;
exercise performance and for control of training. Sports Med 6: 74-7
1996; 22: 157-75 48. Gullstrand L. Physiological responses to short-duration high-
26. Daniels JT, Scardina N, Hayes J, et al. Elite and subelite female intensity intermittent rowing. Can J Appl Physiol 1996; 21:
middle- and long-distance runners. In: Landers DM, editor. 197-209
Sport and elite performers. Champaign (IL): Human Kinetics 49. Edwards RH, Ekelund LG, Harris RC, et al. Cardiorespiratory
Publishers, 1986: 57-72 and metabolic costs of continuous and intermittent exercise
27. Lacour JR, Padilla-Magunacelaya S, Chatard JC, et al. Assess- in man. J Physiol (Lond) 1973; 234: 481-97
ment of running velocity at maximal oxygen uptake. Eur J 50. Fox EL, Bartels RL, Klinzing J, et al. Metabolic responses to
Appl Physiol 1991; 62: 77-82 interval training programs of high and low power output. Med
28. Daniels JT, Yarbough RA, Foster C. Changes in V O2max and Sci Sports Exerc 1977; 9: 191-6
running performance with training. Eur J Appl Physiol 1978; 51. Gorostiaga EM, Walter CB, Foster C, et al. Uniqueness of in-
39: 249-54 terval and continuous training at the same maintained exercise
29. di Prampero PE. The energy cost of human locomotion on land intensity. Eur J Appl Physiol 1991; 63: 101-7
and in water. Int J Sports Med 1986; 7: 55-72 52. Billat
. V, Slawinski J, Bocquet V, et al. Intermittent
. runs at
30. Lger L, Boucher R. An indirect continuous running multistage vV O2max enables subjects to remain at V O2max for a longer
field test: the Universit de Montral Track Test. Can J Appl time than submaximal runs. Eur J Appl Physiol 2000; 81:
Sports Sci 1980; 5: 77-84 188-96

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)
Aerobic Interval Training and Performance 31

53. Brooks GA, Fahey TD, White TP. Exercise physiology: human 75. Casaburi R, Storer TW, Ben-Dov I, et .al. Effect of endurance
bioenergetics and its application. 2nd ed. Mountain View (CA): training on possible determinants of V O2 during heavy exer-
Mayfield Publishing, 1996: 191-5 cise. J Appl Physiol 1987; 38: 1132-9
54. Newman EV, Dill DB, Edwards HT, et al. The rate of lactic acid 76. Martin DE, Coe PN. Better training for distance runners. Cham-
removal in exercise. Am J Physiol 1937; 118: 457-62 paign (IL): Human Kinetics, 1997
55. Belcastro AN, Bonen A. Lactic acid removal rates during con- 77. Babineau C, Lger L. Physiological response of 5/1 intermittent
trolled and uncontrolled recovery exercise. J Appl Physiol aerobic exercise and its relationship to 5 km endurance per-
1975; 39: 932-6 formance. Int J Sports Med 1997; 18: 13-9
56. Newsholm EA. Application of principles of metabolic control 78. Franch J, Madsen K, Djurhuus MS, et al. Improved running
to the problem of metabolic limitations in sprinting, middle economy following intensified training correlates with re-
distance, and marathon running. Int J Sports Med 1986; 7: duces ventilatory demands. Med Sports Sci Exerc 1998; 30:
66-70 1250-6
57. Gimenez M, Servera E, Saunier C, et al. Square-wave endur- 79. Davies CTM, Knibbs A. The training stimulus. The effects of
ance exercise (SWEET) for training and assessment in trained intensity, duration, and frequency of effort on maximum aer-
and untrained subjects. Eur J Appl Physiol 1982; 49: 369-77 obic power output. Int Z Angew Physiol 1971; 29: 299-305
58. Fox EL, Bartels RL, Billing CE, et al. Frequency and duration 80. Henriksson J, Reitman JJ. Quantitative measures of enzyme
of interval training programs and changes in aerobic power. activities in type I and type II muscle fibres of man after
J Appl Physiol 1975; 38: 481-4 training. Acta Physiol Scand 1976; 76: 891-4
59. Demarie S, Koralsztein JP, Billat V. Time limit and time at 81. Tabata I, Nishimura K, Kouzaki M, et al. Effects of moderate-
. intensity endurance and high-intensity intermittent training
V O2max during a continuous and an intermittent running. J .
Sports Med Phys Fitness 2000; 40 (2): 96-102 on anaerobic capacity and V O2max. Med Sci Sports Exerc 1996;
60. Fox EL, Robinson S, Wiegman DL. Intensity and distance of 28: 1327-30
interval training programs and changes in aerobic power. Med 82. Tabata I, Irisawa K, Kouzaki M, et al. Metabolic profile of high
Sci Sports Exerc 1969; 27: 174-8 intensity intermittent exercises. Med Sci Sports Exerc 1997;
61. Anderson O. To optimize your performance, train A la Ver- 29: 390-5
onique. Running Res News 1994 Nov-Dec: 1-3 83. Billat V, Pinoteau J, Petit. B, et al. Reproducibility of running
time to exhaustion at V O2max in sub-elite runners. Med Sci
62. Sjdin B, Jacobs I. Onset of blood lactate accumulation and
Sports Exerc 1994; 26, 254-7
marathon running performance. Int J Sports Med 1981; 2:
84. Womack CJ, Davis SE, Blumer JL, et al. Slow component of
O2 uptake during heavy exercise: adaptation to endurance
63. Robinson DM, Robinson SM, Hume PA, et al. Training inten-
training. J Appl Physiol 1993; 79: 838-45
sity of elite male distance runners. Med Sci Sports Exerc
85. Norris SR, Petersen SR. Effects of endurance training on tran-
1991; 23: 1078-82
. sient oxygen uptake responses in cyclists. J Sports Sci 1998;
64. Billat V, Flechet B, Petit B, et al. Interval training at V O2max: 16: 733-8
effects on aerobic performance and overtraining markers. 86. Burke J, Thayer R, Belcamino M. Comparison of effects of two
Med Sci Sport Sci Exerc 1999; 31: 156-63 interval training programmes on lactate and ventilatory thresh-
65. Smith TP, McNaughton LR, .
Marshall KJ. Effect of 4-wk train- olds. Br J Sp Med 1994; 28: 276-2
ing using Vmax /Tmax on V O2max and performance in athletes. 87. Noakes TD, Myburgh KH, et al. Peak treadmill running veloc-
Med Sci Sports Exerc 1999; 31: 892-6 .
ity during the V O2max test predicts running performance. J
66. Moritani T, Nagata A, De Vries HA, et al. Critical power as a Sports Sci 1990; 8: 35-45
measure of physical working capacity and anaerobic thresh- 88. Fox EL, Mathews DK. Interval Training. Philadelphia (PA):
old. Ergonomics 1981; 24: 339-50 WB Saunders, 1974
67. Gaesser GA, Poole DC. The slow component of oxygen uptake 89. Fox EL, Bartels RL, Billings CE, et al. Intensity and distance
kinetics in humans. Exerc Sports Sci Rev 1996; 24: 35-71 of interval training programs and changes in aerobic power.
68. Poole DC, Ward SA, Gardner GW, et al. Metabolic and respi- Med Sci Sports Exerc 1973; 5: 18-22
ratory profile of heavy and severe exercise. Eur J Appl Phys- 90. Overend TJ, Paterson DH, Cunningham DA. The effect of in-
iol 1988; 31: 1265-79 terval and continuous training in on the aerobic parameters.
69. Whipp BJ. The slow component of O2 uptake kinetics during Can J Sports Sci 1992; 17: 129-34
heavy exercise. Med Sci Sports Exerc 1994; 26: 1319-26 91. Dempsey JA, Hanson PG, Henderson KS. Exercise-Induced
70. Billat V, Binsse .V, Haouzi P, et al. High. level runners are able arterial hypoxemia in healthy human subjects at sea level. J
to maintain a V O2 steady-state below V O2max in an all-out run Physiol (Lond) 1984; 355: 161-75
over their critical velocity. Arch Physiol Bioch 1998; 107: 1-8 92. Katayama K, Sato Y, Ishida K, et al. The effects of intermittent
71. McLellan TM, Cheung SY. A comparative evaluation of the exposure to hypoxia during endurance exercise training on
individual anaerobic threshold and the critical power. Med the ventilatory response to hypoxia and hypercapnia in hu-
Sci Sports Exerc 1992; 24: 543-50 mans. Eur J Appl Physiol 1998; 63: 101-7
72. Roston WL, Whipp BJ, Davis JA, et al. Oxygen uptake kinetics 93. Dempsey JA. Is the lung built for exercise? Med Sci Sports
and lactate concentration during exercise in humans. Am Rev Exerc 1986; 18: 143-55
Respir Dis 1987; 135: 1080-4
73. Henson LC, Poole DC, Whipp BJ. Fitness as a determinant of
oxygen uptake response to constant-load exercise. Eur J Appl
Correspondence and offprints: Vronique Billat, Centre de
Physiol 1989; 59: 21-8
74. Gerbino A, Ward S, Whipp BJ. Effects of prior exercise on Mdecine du Sport CCAS, 2 av. Richerand 75010, Paris,
pulmonary gas-exchange kinetics during high-intensity exer- France.
cise in humans. J Appl Physiol 1996; 80: 99-107 E-mail: veronique.billat@wanadoo.fr

Adis International Limited. All rights reserved. Sports Med 2001; 31 (1)